Special Report, Part 2: The AU’s Less Coercive Diplomacy on Burundi

A group protests the supposed meddling of Rwanda in the internal affairs of Burundi during its ongoing political crisis. Bujumbura, Burundi, February 13, 2015. (Stringer/AFP/Getty Images)

Burundi’s political crisis continues, but it has entered a new phase with the conclusion of the 26th African Union summit on January 31st in Addis Ababa. In a December 2015 Global Observatory article, I analyzed the AU’s novel use of coercive diplomacy in Burundi. This approach came under scrutiny at the January summit, to the point that many consider it a failure. The truth is more complicated.

Before the AU summit, the last decision on Burundi taken by the AU Peace and Security Council (PSC) was set out in its communiqué of December 17th last year. Among other things, this seven-page document authorized the deployment of a 5,000-strong African Protection and Prevention Mission in Burundi (MAPROBU). It gave President Pierre Nkurunziza’s government 96-hours to consent to MAPROBU’s deployment, called for the relaunch of the inter-Burundi dialogue between the government and opposition, and for the complete deployment of the 100 human rights and military observers that the AU had authorized in May 2015. The principal goals of the PSC in taking this decision had been to facilitate a political settlement to Burundi’s ongoing crisis and reduce the threat of armed conflict and violence against civilians.

Moving forward, the PSC’s unprecedented decision left the AU with six options:

  1. It could boost the number of human rights and military observers in Burundi. The AU’s observers started arriving in Bujumbura in late July 2015 but not all of them were able to deploy (moreover, those on the ground were reportedly being hampered by the Burundi government, which wanted to compose their reports jointly with them).
  2. The AU could impose sanctions on Burundi under Article 23(2) of the AU Constitutive Act, which permits it to sanction “any Member State that fails to comply with the decisions and policies of the Union.”
  3. The AU could also choose to continue its coercive diplomacy to secure the Burundi government’s consent for MAPROBU with the mandate set out in the December 17th communiqué (para.13.a.ii).
  4. Alternatively, the AU could continue coercive diplomacy to secure the Burundi government’s consent for MAPROBU but offer an inducement in the form of altering the mission’s mandate from that set out in the communiqué. This represents a “MAPROBU-light” option.
  5. The AU Assembly could consider authorizing a forcible intervention by invoking Article 4(h) of the AU Constitutive Act, which grants the AU the right to intervene militarily under “grave circumstances,” namely, war crimes, genocide, and crimes against humanity. Existing international law on the use of force would require such an intervention to be authorized by a UN Security Council resolution under Chapter VII of the UN Charter.
  6. Finally, the AU could come up with some other form of action.

On December 21st, however, Burundi’s parliament unanimously rejected the proposed AU peacekeeping force. President Nkurunziza then followed suit, stating that any military intervention by AU troops would constitute “an attack on the country and every Burundian will stand up and fight against them.” Was this bombast or a credible threat? Would Nkurunziza really have become the first African president to fight against AU forces? If so, would the Burundian army have remained united after receiving such an order or would it fracture? And could Burundi really keep fighting for the AU in Somalia while fighting against the AU at home? We are unlikely to ever know the answer to these questions. In any event, two days later, Burundi’s foreign minister wrote a letter to the Chairperson of the AU Commission, in which he accused Rwanda of stoking rebellion in Burundi by arming and training refugees, and rejected MAPROBU as tantamount to an invasion force.

It appears that part of the reason why Nkurunziza’s government rejected MAPROBU was a concern that it might end up playing a role similar to the UN peacekeeping operation in Cote d’Ivoire, UNOCI, which helped ensure that the incumbent regime of Laurent Gbagbo was removed after a disputed presidential election. It was probably not lost on Nkurunziza that Gbagbo’s trial at the International Criminal Court started on January 28th, the day before the PSC was scheduled to meet at heads of state level to discuss Burundi.

These developments meant that the AU faced an unprecedented test of its peace and security architecture, which for the first time now involved a threat of military intervention under Article 4(h) of the AU Constitutive Act. Below are some of the key questions raised by the debates and decisions on Burundi taken at the AU summit.

What Decisions on Burundi Were Taken at the AU Summit?

The first relevant decision took place on January 28th, when the AU Executive Council oversaw Burundi’s re-election to serve another two-year term on the PSC. Along with Chad, it was Central Africa’s uncontested choice for that two-year seat. This was an important barometer of the opinion of Central Africa’s governing elites, but also a reflection of the limited pool of high quality candidates for this region.

The next important meeting took place on January 29th, when the PSC met at the level of heads of state and government (the December 17th meeting had convened at ambassadorial level). The meeting considered three issues: terrorism in Africa, and the crises in South Sudan and Burundi. UN Secretary-General Ban Ki-moon also attended the open session. On the authorization of MAPROBU, he said the PSC had “sent a strong signal to the entire continent and the world that you will not stand by while the violence escalates and human rights abuses continue unpunished. I commend your decisive leadership.”

Most member states on the PSC, however, took a different view. The majority of them did not deem it appropriate to send troops to Burundi without the government’s consent. They also agreed it was prudent not to force the issue. Gambia’s President Yahya Jammeh and Tanzania’s Foreign Minister Augustine Mahiga made particularly strong statements against the need for MAPROBU, but they reflected a majority view in the room. Indeed, before signing onto an AU press release supporting MAPROBU on January 8th, Mahiga had publicly broken ranks with the PSC’s December 17th decision not once, but twice, arguing that his country preferred a political settlement over the deployment of troops, which he believed would escalate the current crisis.

On February 6th, after more than a week of additional discussion between senior AU Commission personnel and the PSC members, the AU released the final text of the Council’s decision. Having taken note of Burundi’s rejection of MAPROBU, the heads of state at this meeting emphasized the importance of continuing the inter-Burundi dialogue under East African Community mediation led by Uganda but with AU support. The PSC decided “not to deploy MAPROBU because it considers it premature to send such a force to Burundi.” Instead, the Council decided to dispatch a high-level delegation “to hold consultations” with the Burundi government “as well as other stakeholders…on the inclusive Inter-Burundian Dialogue.”

The composition of the high-level delegation was announced in a press release two days before the AU released the final text of the PSC decision. That statement said the AU Assembly had decided the delegation would comprise five heads of state representing Africa’s five regions: Mauritania (North), South Africa (Southern), Senegal (West), Gabon (Central), and Ethiopia (Eastern). None of these leaders have been in power as long as President Nkurunziza. Significantly, the AU press release expanded the delegation’s mandate from a sole focus on the inter-Burundi dialogue (as expressed in the PSC communiqué of the January 29th meeting), to also include “the deployment of the African Prevention and Protection Mission in Burundi (MAPROBU), if accepted by the Government of Burundi.”

The main event of the summit was the AU Assembly meeting, which convened January 30-31st (more than two weeks on, the final decision from the meeting has not been released due to differences between the initial proposed text and the adopted version, but its general content is now well-known). The majority of the Assembly consultations took place in closed sessions involving heads of state and their small entourages. In order to authorize an Article 4(h) military intervention in Burundi, the Assembly would have needed the support of two-thirds of AU member states. In fact, the Assembly decided the conditions in Burundi did not warrant such a response. Instead, they dispatched the high-level delegation noted earlier. This mirrored the approach the AU had adopted with regard to the uprisings in Libya in early 2011. In that case, when the rebellion escalated, the UN Security Council rejected the AU’s approach in favor of enforcement action authorized in resolutions 1970 and 1973.

How Credible was the PSC’s Threat of Forcible Intervention?

The threat of an Article 4(h) intervention was always going to prove difficult to implement. In generic terms, the credibility of such threats can be enhanced if they are communicated clearly and consistently (the more actors voicing their support, the better), if the coercer has a reputation for following through, and if they possess the required material capabilities. The AU struggled on each of these criteria.

First, the AU does not have a reputation for carrying out a threat of military enforcement against one of its member states. In addition, it has never before invoked Article 4(h) in such circumstances. The only prior occasion where the AU has threatened and then used military force against an incumbent regime came in March 2008 on the tiny Comoran island of Anjouan. Here, the AU’s Operation Democracy in the Comoros, spearheaded by troops from Tanzania and Sudan, but also involving soldiers from Senegal and logistical support from Libya, forced the incumbent ruler Mohammed Bacar to step down after he had organized an illegal election in order to cling onto power. In the Comoros, the AU used military force to restore constitutional governance. In Burundi, however, the PSC’s threat was made in the name of protecting civilians and preventing the escalation of violence.

Second, the threat was not communicated with complete consistency and clarity because the PSC failed to maintain a public show of unity. As noted earlier, Tanzania’s foreign minister was particularly vocal in departing from the PSC’s stated objective as set out in the December 17th communiqué. Moreover, in the midst of the Assembly meetings, Ibrahim Fall, the AU’s special representative for the Great Lakes region, publicly argued that the AU had never intended to forcibly intervene in Burundi and that the entire concept had been “unimaginable.”

The PSC’s threat was weakened still further because of the complex layers of decision-making involved. The PSC could only recommend that the AU Assembly consider authorizing an Article 4(h) intervention. In addition, even if the AU Assembly had authorized such an intervention, in order to conform with existing international law on the use of force it would have required a UN Security Council resolution passed under Chapter VII of the UN Charter. This is made clear in Article 53(1) of the UN Charter, which sets out the role of regional arrangements in the UN system. This added yet another layer of uncertainty into the decision-making process: would the AU carry out what would probably be seen as an illegal use of force if it did not obtain UN Security Council authorization?

This raises the question of how the Security Council would have reacted in the hypothetical event that the AU Assembly had authorized an Article 4(h) intervention? It is impossible to know for sure but it is interesting to speculate, especially if we assume that this is probably not the last time that the AU will invoke Article 4(h). The only potential precedent at the UN is Security Council Resolution 1973, which authorized the use of air power to protect civilians in Libya. It did not authorize the use of ground troops, which was the AU’s plan for Burundi. Would the Security Council have been willing to pass another, similar resolution on Burundi?

On the one hand, recent Security Council practice would appear to support the idea that if a regional organization called for such action, the Security Council would be more likely to endorse it than if there was no such regional request. In that case, much would have hinged on the position of the “A-3 states” (i.e. the three African non-permanent members of the Security Council), currently Angola, Egypt, and Senegal. However, Angola had supported Nkurunziza’s third term bid and Egypt also emphasized that any peacekeeping deployment should take place with the consent of the host state authorities. Moreover, even if a forcible deployment had the AU’s support, providing UN financial or logistical support in such extraordinary circumstances would have been deeply controversial. Finally, the international controversy over the NATO-led operations in Libya during 2011, particularly its over-stepping the authorized mandate and failing to plan sufficiently for the aftermath of a rebel victory, probably means that there is little likelihood of Resolution 1973 being repeated, and this time with ground troops.

Credible threats also require the material capabilities to carry them out. So, was the AU ready to conduct an Article 4(h) intervention in Burundi? On Burundi, the AU’s contingency planning was led by the Eastern Africa Standby Force (EASF). Declared operational in December 2014, the EASF has yet to deploy its composite regional force but it would probably have been able to generate a force if Burundi’s government had consented to MAPROBU’s deployment. Indeed, the EASF’s contingency planning was based solely on the assumption that Burundi’s government would accept MAPROBU. This position was formalized at an extraordinary summit, of the EASF heads of state on January 30th, which was held immediately before the start of the AU Assembly meeting. Here, eastern African leaders decided “to authorize the deployment of EASF as part of the MAPROBU subject to consent by the Republic of Burundi” (para.3). Similarly, the draft concept of operations for MAPROBU did not even broach the possibility of the EASF conducting a forcible intervention in Burundi.

There is also the issue of logistical support. If the AU Assembly had authorized MAPROBU’s deployment (either with or without the consent of Burundi’s government), it would have needed external donors to assist the mission financially and with logistics support. In its December 17th communiqué, the PSC had explicitly asked for such UN support in the event of a consensual MAPROBU deployment. Presumably such needs would only have increased in the case of forcible intervention. The fact that the UN Department of Peacekeeping Operations conducted its own contingency planning on how best to deploy a UN force to Burundi suggests there was considerable anxiety that the AU might not be able to deploy MAPROBU.

Finally, there is the issue of Burundi’s peacekeepers in the AU mission in Somalia (AMISOM). Here, the AU chose not to use arguably its most important source of potential leverage over the Burundi government: the withdrawal of its forces from AMISOM. These forces are important to the Burundian government for two main reasons. First, these peacekeeping deployments have provided a way to cement the military integration process that occurred after Burundi’s last civil war. Second, they generate significant financial revenues because of the training and assistance packages that several donors provide to facilitate such deployments, notably the United States and France, and the monthly allowances retained by the individual peacekeepers, much of which they invest back home when they complete their tour of duty. In AMISOM, Burundian troops earn approximately $828 USD per month in allowances, paid for by the European Union (the Burundian government retains $200 USD of the monthly allowance for administrative and other purposes). By the end of 2015, Burundian peacekeepers deployed abroad were reportedly the government’s “most important source of foreign exchange.”

Although it would presumably be untenable for the AU to forcibly intervene in Burundi while relying on its peacekeepers in AMISOM, the PSC did not threaten to withdraw Burundi’s contingent from Somalia. It appears that such a course of action was thought to be imprudent because the return of so many soldiers missing their extra income could exacerbate the situation at home. The AU therefore chose not to use its main potential source of leverage over President Nkurunziza’s regime.


Some senior Burundian officials were jubilant in declaring the demise of MAPROBU after the PSC meeting on January 29th. For example, Nkurunziza’s special adviser on communications, Willy Nyamitwe, pronounced MAPROBU dead and even showed the world how he would celebrate. He went on to question why the AU would bother to send a delegation to Burundi when the government’s position was already known.

However, it’s important to recall that the December 17th communiqué authorized MAPROBU for an initial period of six months and neither its subsequent decision on January 29th nor the AU Assembly decision un-authorized the mission. Rather, the PSC subsequently decided its deployment would be premature and the Assembly said that there was no current basis for an Article 4(h) military intervention in Burundi. This means that MAPROBU’s deployment prospects, while significantly diminished, hinge, once again, on a high-level dialogue between the government of Burundi and the AU delegation. Whether there is scope for Burundi to accept a “MAPROBU-light” option (i.e. with a different mandate from the one written into the December communiqué) remains to be seen.

In sum, there might yet be scope for a MAPROBU deployment, either as a result of the efforts of the AU high-level delegation, or to help implement any agreement that might emerge from the inter-Burundi dialogue. In this sense, ensuring real progress in the inter-Burundi dialogue might resurrect MAPROBU, although perhaps not in the form it was originally articulated.

How Should We Assess Events in Burundi Since December 17th?

This chain of events has provoked numerous criticisms of the AU. Even the AU’s incoming chairperson, Chad’s long-time President Idriss Deby, seemed to criticize the AU as a “do nothing” organization. “Our organisation acts as it has for the past 20 or 30 years: we meet often, we talk too much, we always write a lot, but we don’t do enough, and sometimes not all [sic.].”

But in spite of all these problems, it would be wrong to conclude that the PSC’s threat of Article 4(h) intervention had no positive effects. The dominant view in media commentary is a multidimensional victory for the Burundi government: MAPROBU has been killed off; the promised dialogue remains stalled, in part due to government intransigence; sanctions have not materialized; most AU observers were kept out and those on the ground prevented from completing their reports. In sum, Nkurunziza’s regime effectively called the AU’s bluff on Article 4(h) and used astute diplomacy to win over reportedly almost all the members of the PSC at the January 29th summit. After that, there was no chance of two-thirds of the AU Assembly supporting an Article 4(h) intervention, even if the circumstances on the ground in Burundi warranted them. Here, it’s important to note that at the end of January 2016, the situation in Burundi probably did not warrant forcible military intervention on either moral or prudential grounds. Indeed, some commentators suggested the only plausible way to resolve Burundi’s crisis was to stop even talking about intervention and instead fully accept that Nkurunziza should remain in power and hope that he didn’t treat his opponents too badly afterwards.

Although it’s too early to say for sure if the gamble of using coercive diplomacy worked, we should recall what the PSC was trying to achieve when it recommended implementing Article 4(h). It had two principal goals: ideally, facilitate a political settlement to the ongoing crisis, and also de-escalate armed conflict and violence against civilians in Burundi. It is therefore notable that there has not been another episode of attacks and summary executions on the scale of December 11-12th; although reported fatalities since late December have continued at between 15 and 22 per week. It is plausible that the reduced number of killings and the relaunch of the inter-Burundi dialogue is in large part due to the PSCs threat of Article 4(h) intervention and subsequent coercive diplomacy, which also helped gain support for the UN Security Council to visit Burundi in mid-January 2016. An alternative view is that the AU communiqué and subsequent international diplomacy may have provided only a temporary pause in the violence while government and rebel forces waited to see what international actors would do. Indeed, a new rebel group—the Union of Patriots for the Revolution—declared its formation on February 11th. Since the international diplomacy from mid-December did not resolve the underlying disputes, the parties might now be more likely to ramp up the violence with civilians inevitably caught in the middle. For the sake of Burundians and stability across the Great Lakes region, international diplomats inside and outside Africa must continue to work to ensure this does not happen.

Paul D. Williams is Associate Professor of International Affairs at George Washington University. @PDWilliamsGWU.